Freescale Semiconductor MC68881 - Double Precision Binary Real Format

Table 3-2. Double Precision Binary Real Format
Memop/Format:
Field Size (in Bits):
s = Sign
e = Biased Exponent
f = Fraction
Total
Interpretation of Sign:
Positive Mantissa, s =
Negative Mantissa, s =
Normalized Numbers:
Bias of e
Range of e
Range of f
Mantissa = Significand =
Relation to Representation of Real Numbers
Denormalized Numbers:
e = Format Minimum =
Bias of e
Range of f
Mantissa = Significand =
Relation to Representation of Real Numbers
Signed Zeros:
e = Format Minimum =
f = Mantissa = Significand =
Signed Infinities:
e = Format Maximum =
f = Mantissa = Significand =
NANs (Not-A-Number):
S =
e = Format Maximum =
f=
Representation of f
XXXX... XXXX
f When Created by the FPCP
Ranges (Approximate):
Maximum Positive Normalized
Minimum Positive Normalized
Minimum Positive Denormalized
63 62 52 51
S I BIASED I
EXPONENT
FRACTION
+ 1023
0 < e < 2047 ($7FF)
Zero or Nonzero
1.f
(-1)s x 2e-1023 x l.f
0 ($OOO)
+ 1022 ($3FE)
Nonzero
g.f
(-1Is x 2-1022 x O.f
0 ($ool
O.f = 0.0
2047 ($7FF)
O.f = 0.0
DON't Care
2047 ($TFF)
Nonzero
0.1xxxx... xxxx, Nonsignaling
O.Oxxxx... xxxx, Signaling
Nonzero Bit Pattern
.11111 ,.. 1111
18 x 10307
2.2 x 10 =308
4.9 x 10 -324
0
I
MC68881/MC68882 USER'S MANUAL FREESCALE
3-11